Injection of a cell lineage tracer (HRP) into Drosophila embryos before cellularization provides a way of selectively labelling cells at later stages that have undergone only a few mitoses. All cells born and differentiating during embryogenesis become labelled, whereas further proliferation and growth during postembryonic development causes an almost complete dilution of the marker in the adult cell complement. Early born neurons visualized in this way are good candidates for executing a pioneering function during postembryonic differentiation of the adult nervous system.

In all three pairs of leg imaginal discs, a stereotyped set of larval sense organs becomes selectively labelled. Their axons fasciculate with a larval nerve, which connects the leg disc with the central nervous system. Larval sense organs are not present in the other imaginal discs.

Larval neurons are not present in the transformed antennal discs of Antp 73B flies. Nonetheless adult axons successfully navigate to the base of these discs as they differentiate to form ectopic legs. We conclude that embryonically formed larval nerves are not essential for the guidance of adult axons within the leg discs.

In holometabolous insects, the imaginal appendages develop from nests of undifferentiated cells, the imaginal discs. Growth and differentiation of the discs mainly occurs in the late larva and in the pupa. Thus, the adult sensory structures are formed during these stages of postembryonic development. The newly growing sensory axons are faced with the problem of finding their way to the CNS and establishing specific synaptic contacts with their target cells. The tendency of growing neuronal fibers to fasciculate with other fibers and to follow their pathways is well known (e.g. Bate, 1976, 1978; Bentley and Caudy, 1983; Berlot and Goodman, 1984; Bastiani et al. 1985; Kuwada, 1986). Accordingly, the pathways laid down by larval neurons during embryogenesis may provide a matrix for the proper orientation of newly growing imaginal axons during postembryonic development of holometabolous insects.

Using monoclonal antibodies that recognize the peripheral nervous system in Drosophila, the formation of adult neuronal pathways has previously been described for different imaginal discs (Murray et al. 1984; Palka, 1986; Jan et al. 1985). In the leg discs, a set of larval neurons has also been identified. It has been argued that the pathways established by these neurons provide an essential source of guidance for the newly growing adult sensory axons (Jan et al. 1985).

A simple method allows early differentiated larval cells to be identified within developing imaginal tissues that have undergone further profiferation and growth during postembryonic development. Applying this method we have previously identified a small set of larval neurons in the optic lobes of the brain (Tix et al. 1989). Here we use the same approach for a screen of larval cells in the growing leg imaginal discs. In addition to labelling with HRP the neurons previously described (Jan et al. 1985), we were able to identify the sheath cells associated with them and to inspect the structure and projection pattern of the larval components in the leg imaginal discs more closely. At the same time, we have looked at the pattern of nerve pathways formed in the ectopic leg discs of Antennapedia 73B flies. These discs do not contain larval nerves and we have used them to test the idea that larval nerves are essential for the guidance of adult axons.

Stocks

Oregon R wildtype and mutant Antp 73B flies of Drosophila melanogaster were used.

Labelling embryonic cells with HRP

Embryos at the syncytial blastoderm stage or at early stages of cellularization were dechorionated, fixed on a slide, desiccated and covered with 10S fluorocarbon oil (Voltalef). 5–15 nl (depending on the degree of egg desiccation) of 7.5 % or 10 % HRP in 0.2 M-KCI was injected centrally into the yolk at ca. 50% EL (egg length). The tracer spreads throughout the embryo and will then become incorporated into all the cells during cell closure. Since a lineage tracer like HRP does not pass between cells because of its size, it becomes trapped in each of the cells as soon as the last cytoplasmic bridges between the somatic cells are degraded during the first minutes of gastrulation. For further details see Technau (1986).

Following injection the globally labelled embryos were allowed to continue their development under 3S fluorocarbon oil (Voltalef) in a humid chamber at 25 °C. 18–20 hours after the injection, first instar larvae hatched from the eggs and were collected on agar dishes supplemented with yeast. The individuals were raised until they reached a given stage of postembryonic development, varying between the second larval instar and early pupal stages.

Histology

Larvae or pupae were dissected in 0.1 M-phosphate buffer (PB, pH 7.2) and the imaginal discs of the eyes and antennae, legs, wings and halteres were removed. The organs were fixed in 2.5% glutaraldehyde in PB for 30min, incubated for 20-30 min in diaminobenzidine staining solution (1mgml−1 DAB, 0.003% H2O2, 0.1M-PB), washed in PB, dehydrated and embedded in Epon. For light microscopy, the preparations w ere either embedded as whole mounts on slides or cut into semithin sections (1.5 μm) which were counterstained with a solution of 0.05 % methylene blue, 0.01 % toluidine blue and 0.05 % borax at 60°C. For electron microscopy, the specimens were treated as for light microscopy except that following the reaction with the DAB staining solution they were washed and transferred to 2% OsO4 for 2h, dehydrated, embedded and cut into pale gold ultrathin sections. Sections were stained with 1.5 % aqueous lead citrate (pH 12) for 15 min.

Immunohistochemistry

MAb 22C10 has been isolated during a search for MAbs against the Drosophila nervous system (Fujita et al. 1982) and has been kindly provided to us by A. Ferrus. The 22C10 antigen is located in neuronal cellular membranes and can be visualized from the initial moment of axon growth. The staining decays in the soma when the neuron matures but remains in the axon during further development (Canal and Ferrus, 1986). In other experiments, we used an anti-peroxidase antibody (Jan and Jan, 1982) to stain larval and adult nerves in imaginal discs.

CNS and imaginal discs were removed in PBS and fixed for 1h in 4% paraformaldehyde (in PBS). The antibody staining was performed according to the protocol of Jan and Jan (1982). The stained preparations were dehydrated in an ethanol series and embedded in Epon as whole mounts.

Illustrations

Labelled cells in the whole-mount preparations were photographed using Nomarski optics or drawn by use of a camera-lucida device (Zeiss). Electron micrographs were taken with a Siemens Elmiskop 101.

During embryogenesis the larval progenitor cells perform only few postblastodermal mitoses, and the cells decrease in size following each division. During post-embryonic development, however, the adult progenitor cells undergo a large number of divisions and there is considerable cell growth. Accordingly, a lineage tracer like HRP incorporated into all the cells during cellularization will label the entire larval cell lineage (Technau, 1986) whereas in most of the adult tissues the tracer becomes diluted to such a degree that it is no longer detectable. This dilution is especially evident in certain regions of the brain and thoracic ganglia and in the imaginal discs.

Larval sense organs in the leg imaginal discs

After injection of HRP according to the method described above, all imaginal disc cells are labelled until the 2nd larval stage, reflecting the fact that proliferative activity in the discs is still low at this stage. In third instar larvae massive proliferation and growth leads to an almost complete dilution of the marker. In the leg imaginal discs, the marker becomes diluted in all cells except for a small population of cells that still carry a high concentration of HRP (Fig. 1). Thus, these cells, unlike their neighbours, have not resumed dividing, and the label persists.

Fig. 1.

Whole-mount preparations (A, C–E) and camera-lucida drawing (B) showing the distribution of larval sensory organs and neuronal pathways in wildtype third instar leg discs. (A–C), sagittal view; (D,E), horizontal view. Injection of HRP in early embryos uncovers a stereotyped set of neurons and non-neuronal support cells in all three pairs of leg discs. They represent sense organs located in the presumptive tarsal epithelium (S1–S3; A–D) and a single neuron (S4) with a long dendritic process lying superficially above the presumptive tibia-femur region (B,E). Sensory organs S2 and S3 have long dendritic processes (B,D) which innervate the Keilin organs. The axons of all neurons converge into a larval nerve, which is also labelled and which connects the leg disc and the corresponding thoracic ganglion (A–C). The distribution of sensory organs disclosed following HRP-treatment (A,B) corresponds to the distribution of neurons detected by MAb 22C10 (C). Ad, adepithelial cells; BM, basement membrane; D, long dendrites; Ep, disc epithelium; HS, hypodermal stalk; L, presumptive leg lumen; LN, larval nerve; NS, nerve stalk; P, peripodial membrane; S1-S3, tarsal sensory organs; S4, nonepithelial neuron; Ta, presumptive tarsal region; Tr, trachea. (A,C,E): composite photographs. Bar, (A,C–E) 20μm.

Fig. 1.

Whole-mount preparations (A, C–E) and camera-lucida drawing (B) showing the distribution of larval sensory organs and neuronal pathways in wildtype third instar leg discs. (A–C), sagittal view; (D,E), horizontal view. Injection of HRP in early embryos uncovers a stereotyped set of neurons and non-neuronal support cells in all three pairs of leg discs. They represent sense organs located in the presumptive tarsal epithelium (S1–S3; A–D) and a single neuron (S4) with a long dendritic process lying superficially above the presumptive tibia-femur region (B,E). Sensory organs S2 and S3 have long dendritic processes (B,D) which innervate the Keilin organs. The axons of all neurons converge into a larval nerve, which is also labelled and which connects the leg disc and the corresponding thoracic ganglion (A–C). The distribution of sensory organs disclosed following HRP-treatment (A,B) corresponds to the distribution of neurons detected by MAb 22C10 (C). Ad, adepithelial cells; BM, basement membrane; D, long dendrites; Ep, disc epithelium; HS, hypodermal stalk; L, presumptive leg lumen; LN, larval nerve; NS, nerve stalk; P, peripodial membrane; S1-S3, tarsal sensory organs; S4, nonepithelial neuron; Ta, presumptive tarsal region; Tr, trachea. (A,C,E): composite photographs. Bar, (A,C–E) 20μm.

In all three pairs of leg discs, the persistent labelled cells can be found in the same stereotyped pattern. Three prominent clusters of labelled cells are located in the presumptive tarsal region of the disc epithelium (according to the fate map of Schubiger, 1968). These clusters include neurons and non-neuronal support cells and thus represent fully differentiated larval sense organs. We will refer to these sense organs as S1-S3. SI comprises three, S2 four and S3 nine labelled cells in the disc epithelium. The nine cells in S3 are arranged in three closely attached subunits of three cells each. S2 possesses one and S3 three very long dendrites (Fig. 1B,D), which run between the apical face of the epithelium and the peripodial membrane and project through the stalk to the larval hypodermis in order to innervate the Keilin’s organs, which are located in the ventrolateral epidermis of each thoracic segment. Besides its long dendrite, S2 possesses a short dendritic element immediately terminating at the apical face of the disc epithelium. SI only shows one dendritic specialization which is of the same type as the short one in S2. In addition to these sensory organs in the presumptive tarsal region, a single neuron (S4) with a large cell body exists, which is attached to the surface of the disc epithelium in the presumptive tibia/femur region (Fig. 1B,E). Its dendrite is enveloped by a prominent sheath that stretches along the surface of the disc and parts from it at a place other than the hypodermal stalk to project towards the larval epidermis. Its axon and the axons of the sensory neurons of S1-S3 project towards the basement membrane of the disc to fasciculate with a larval nerve (carrying 70–80 axons; Fig 2A,C), which proceeds along the disc and connects it with the respective thoracic neuromere of the CNS (Fig. 2B). It is likely that S1-S3 become internalized together with the progenitor cells of the leg discs. As a result, the S2- and S3-dendrites elongate and remain in contact with the larval epidermis via the hypodermal stalk to innervate the Keilin’s organs whereas S4 may get secondarily attached to the disc.

Fig. 2.

Whole-mount preparations (A,B) and electron micrographs (C–F) showing details of HRP-labelled structures in third instar leg imaginal discs. (A) Axons of the larval sense organs fasciculate with a larval nerve (arrow), which passes the imaginal leg disc beyond the basement membrane. Labelled cells near the nerve presumably are adepithelial cells. (B) The sensory axons follow this larval nerve through the nerve stalk into the respective thoracic neuromere of the ventral cord. On the way to the ventral cord a second larval nerve (arrow) joins the stalk. (C) Ultrathin cross section through the larval nerve at the base of the disc carrying ca. 80 axons. (D) Ultrathin section cut in the centre of the disc showing a cross section through the non-epithelial basal part of labelled sense organ S3 and a closely associated adepithelial cell (arrowhead) which is enwrapped by an HRP-labelled support cell of S3. Compare unlabelled adepithelial cell in the disc lumen (arrow). (E) Ultrathin cross section through epithelial part of sense organ S3 in the presumptive tarsal region revealing a triscolopidial organization. Arrows point to three subunits each composed of a dendrite which is enwrapped by non-neuronal support cells. (F) Longitudinal ultrathin section through distal epithelial region of HRP-labelled sense organ SI showing outer and inner segment of dendrite and surrounding non-neuronal support cells. Ad, adepithelial cells; BM, basement membrane; Ep, disc epithelium; ID, inner dendritic segment; LN, larval nerve; NS, nerve stalk; OD, outer dendritic segment; SI,S3, tarsal sensory organs; SC, non-neuronal support cells; TG, thoracic ganglia. Bar, (A,B) 20μm; (D) 5μm; (C,E,F) 2μm.

Fig. 2.

Whole-mount preparations (A,B) and electron micrographs (C–F) showing details of HRP-labelled structures in third instar leg imaginal discs. (A) Axons of the larval sense organs fasciculate with a larval nerve (arrow), which passes the imaginal leg disc beyond the basement membrane. Labelled cells near the nerve presumably are adepithelial cells. (B) The sensory axons follow this larval nerve through the nerve stalk into the respective thoracic neuromere of the ventral cord. On the way to the ventral cord a second larval nerve (arrow) joins the stalk. (C) Ultrathin cross section through the larval nerve at the base of the disc carrying ca. 80 axons. (D) Ultrathin section cut in the centre of the disc showing a cross section through the non-epithelial basal part of labelled sense organ S3 and a closely associated adepithelial cell (arrowhead) which is enwrapped by an HRP-labelled support cell of S3. Compare unlabelled adepithelial cell in the disc lumen (arrow). (E) Ultrathin cross section through epithelial part of sense organ S3 in the presumptive tarsal region revealing a triscolopidial organization. Arrows point to three subunits each composed of a dendrite which is enwrapped by non-neuronal support cells. (F) Longitudinal ultrathin section through distal epithelial region of HRP-labelled sense organ SI showing outer and inner segment of dendrite and surrounding non-neuronal support cells. Ad, adepithelial cells; BM, basement membrane; Ep, disc epithelium; ID, inner dendritic segment; LN, larval nerve; NS, nerve stalk; OD, outer dendritic segment; SI,S3, tarsal sensory organs; SC, non-neuronal support cells; TG, thoracic ganglia. Bar, (A,B) 20μm; (D) 5μm; (C,E,F) 2μm.

On their way towards the larval nerve, the axons of S2/S3 and SI respectively establish two separate pathways within the disc (Fig. 1A,B). On the S2/S3-pathway the axons and their enveloping sheath cells meet two single cells located in the lumen of the prospective tarsus. As judged from their shape, these cells are adepithelial cells which are supposed to give rise to adult muscles (Poodry and Schneiderman, 1970). Electron microscopy reveals that these cells are enwrapped by an HRP-labelled sheath, which presumably belongs to glial components of the adjoining sensory organs (Fig. 2D). Another 4-7 of these supposed adepithelial cells occur near the place where the disc sensory axons fasciculate with the larval nerve (Fig. 1A,B; 2A).

MAb 22C10 labels the neuronal components whereas the non-neuronal support cells remain unlabelled (Fig. 1C). In third instar leg discs, the distribution of labelled cells essentially agrees with the previously described pattern using another neuron-specific antibody (MAb 21A4; Jan et al. 1985), and it corresponds to the location of the sensory organs described above. Since MAb 22C10 only incompletely stains the neuronal cell bodies, we were not able to determine the precise number of neurons in these preparations. But the number of HRP-labelled dendritic elements described above and electron microscopy suggest that there are seven sensory neurons in all, three pairs of leg discs in third instar larvae, three in S3 (Fig. 2E), two in S2 and one each in SI (Fig. 2F) and S4. This is in agreement with the number of neurons found by Jan et al. (1985) by staining with MAb 21A4.

During the first few hours after puparium formation (apf) the leg discs evaginate. Four hours apf they have already assumed a tubular shape and individual segments become distinguishable (Fig. 3A,B). The HRP-labelled sensory organs S1-S3 have moved with the tip of the leg. Their axons have elongated and the two pathways formed by these axons are clearly visible in the lumen of the leg down to the base where they join the larval nerve on its way to the CNS. The single large neuron S4 remains located near the base of the leg in the femoral region close to the larval nerve. We do not know whether the labelled structures persist through metamorphosis since secretion of the cuticle prevents the histochemical demonstration of HRP at later pupal stages.

Fig. 3.

Camera-lucida drawing (A) and whole-mount preparation (B) of an evaginating leg 5h and 4h after puparium formation, respectively. During evagination of the leg, the axons of sensory organs S1–S3 elongate running from the tip of the tarsus through the leg tube down to the base where they enter the larval nerve. They form two pathways (S2/3 and SI) in the anterior and the posterior half of the leg lumen. Ad, presumed adepithelial cells; LN, larval nerve; S1–S4, larval sensory organs. Ta, tarsus. Bar, (B) 30 μm.

Fig. 3.

Camera-lucida drawing (A) and whole-mount preparation (B) of an evaginating leg 5h and 4h after puparium formation, respectively. During evagination of the leg, the axons of sensory organs S1–S3 elongate running from the tip of the tarsus through the leg tube down to the base where they enter the larval nerve. They form two pathways (S2/3 and SI) in the anterior and the posterior half of the leg lumen. Ad, presumed adepithelial cells; LN, larval nerve; S1–S4, larval sensory organs. Ta, tarsus. Bar, (B) 30 μm.

We also inspected other imaginal discs of HRP-labelled individuals for the presence of early differentiated cells with a low rate of proliferation. Neither in the wing and haltere discs nor in the eye/antennal discs did we find any differentially labelled cells. In these discs the label becomes equally diluted in all cells during second and third instar larval stages. Thus, all cells of these discs undergo a large number of postembryonic mitoses. This is in agreement with the fact that in the late third larval instar neither MAb 21A4 (Jan et al. 1985) nor MAb 22C10 detects any larval neuronal cells in these discs.

Nerve pathways in the discs of Antennapedia flies

For comparison, we used an anti-HRP antibody to stain and investigate the pattern of nerves that develops in the transformed antennal discs of Antp 73B flies. In such flies, the imaginal disc in the antennal position differentiates at metamorphosis to form an ectopic leg. As expected, the antibody staining reveals no larval neurons in the disc of early third instar larvae (Fig. 4). Shortly before pupariation, however, neurons begin to differentiate in the distal disc epithelium and, as evagination begins, these neurons send axons proximally to the base of the differentiating leg disc. Despite the absence of preformed larval pathways, these newly differentiating nerves grow directly to the base of the disc, usually, but not always, forming the distinct axon pathways which are comparable to the pathways formed in normal leg discs (Fig. 4).

Fig. 4.

Eye (ed) and ectopic leg (el) discs from Antp 73B larvae and pupae, stained with an antibody against HRP to reveal associated nerve pathways. (A) Eye and evaginating leg discs 12 hours after pupariation (25°C). Two nerve pathways run down the lumen of the left-hand leg, although only one (arrowed) is visible in this plane of focus. Similar pathways (not in focus) are present in the right-hand disc. (B) Two pathways are not always formed in the leg discs; here a single nerve bundle runs from distal to proximal in the lumen of the ectopic leg (12 h after pupariation). (C) Eye and prospective leg disc (pld) from an early third instar Antp 73B larva. There are no neurons in the prospective leg disc at this stage. (D) The same at a more superficial plane of focus to show staining of the larval photoreceptor nerve (Bolwig’s nerve; arrowheads) as it passes over the surface of the leg and eye discs, ed, eye disc; el, ectopic leg disc; pld, prospective leg disc. Bar, (A–D) 25 μm.

Fig. 4.

Eye (ed) and ectopic leg (el) discs from Antp 73B larvae and pupae, stained with an antibody against HRP to reveal associated nerve pathways. (A) Eye and evaginating leg discs 12 hours after pupariation (25°C). Two nerve pathways run down the lumen of the left-hand leg, although only one (arrowed) is visible in this plane of focus. Similar pathways (not in focus) are present in the right-hand disc. (B) Two pathways are not always formed in the leg discs; here a single nerve bundle runs from distal to proximal in the lumen of the ectopic leg (12 h after pupariation). (C) Eye and prospective leg disc (pld) from an early third instar Antp 73B larva. There are no neurons in the prospective leg disc at this stage. (D) The same at a more superficial plane of focus to show staining of the larval photoreceptor nerve (Bolwig’s nerve; arrowheads) as it passes over the surface of the leg and eye discs, ed, eye disc; el, ectopic leg disc; pld, prospective leg disc. Bar, (A–D) 25 μm.

The technique we use to detect larval cells in developing imaginal tissues is a simple one. A lineage tracer (HRP) is incorporated into all cells of the early embryo and the marker becomes differentially diluted in the progeny cells by a varying number of subsequent cell divisions during embryonic and postembryonic development.

Screening leg imaginal discs of third instar larvae by using this method discloses a stereotyped pattern of HRP-labelled cells in an otherwise unlabelled background in all three pairs of leg discs. Among these cells are 7 neurons that are identical to those previously detected by staining with a neuron-specific antibody (Jan et al. 1985). In addition, our approach allowed us to identify the non-neuronal sheath cells which together with these neurons compose 4 sensory organs as well as a number of apparent adepithelial cells associated with the sensory structures. Furthermore, the concentration of the lineage tracer reflects a low rate of proliferation and growth of these cells. This, and the fact that the sensory organ cells are the only fully differentiated cells in the leg discs, suggests that they are born and differentiate during embryogenesis. Finally, selectively labelling the larval components in the discs with HRP allows their structure and projection patterns to be analysed in detail.

The sense organs are of different morphological types as judged from their cell numbers and their dendritic specializations and may serve different mechanosensory functions. S3 sends three dendrites and S2 one very long dendrite through the hypodermal stalk to the larval epidermis where they innervate Keilin’s organs. Thus, it seems that these larval sensory organs, which became internalized with the leg disc progenitor cells, still perceive external stimuli. A basal sensillum associated with the hypodermis was described by Keilin (1915) for the leg discs of different dipterans. He supposed that these disc sensilla were relics of larval leg receptors conserved in the course of evolution. The dendrite of S4 also seems to remain in contact with the larval epidermis but does not project through the hypodermal stalk. Since S4 is not directly associated with sheath cells and since it is attached to the disc surface rather than being integrated into the disc epithelium like S1-S3, the nature of this structure remains obscure. It could have got attached to the disc after segregation of the disc progenitor cells. Unfortunately, we were not able to determine whether these larval sensilla persist through metamorphosis to become part of the adult peripheral nervous system.

As already stressed by Jan et al. (1985), the pathways formed by the early neurons in the leg discs might serve a pioneering function in guiding towards the CNS the axons of adult sensory neurons that start differentiation after the onset of metamorphosis. Two different populations of larval sensory neurons are involved in establishing the initial pathway to the CNS. The distal part of the pathway within the leg disc is formed by the axons of sensory organs S1-S3. Since these organs reside in the prospective tarsal region their axons elongate during evagination of the disc and run along the entire length of the leg down to the base. Here, they fasciculate with a nerve that connects the disc with the CNS and thus forms the proximal part of the pathway. This nerve is composed of the axons of a second population of larval sensory neurons that reside in the larval epidermis. Thus, developing adult sensory neurons over the entire surface of the leg have to send their axons only a short distance in order to reach and fasciculate with one of the two parallel tracts formed by the larval sensory organs S1–S3 (see also Jan et al. 1985).

However, it is unlikely that the role of the larval neurons associated with the leg discs is simply to pioneer the subsequent route of the adult axons that grow out of the disc epithelium when it differentiates at metamorphosis. These neurons have a function as part of the normal sensory equipment of the larva, and it may be incidental to their function as sensory cells that they are incorporated into the invaginated disc epithelium. Because of their position at the distal tip of the disc, their axons are passively extended over the length of the leg as it evaginates and they provide a pathway to the base of the disc for any axons that subsequently grow out in the differentiating leg. However, the evidence of successful distal to proximal axon growth in the ectopic leg discs of Antp 73B flies suggests that the presence of an embryonically formed larval pathway is not essential for the distal to proximal guidance of adult axons in evaginating discs. In other discs our screen failed to reveal cells with a low number of cell divisions and we do not find any larval neurons in these discs when we stain them with anti-HRP antibodies. The conclusion that larval nerves are not required for adult nerve guidance within imaginal discs is supported by experiments with the wing disc. These show that the normal navigation of differentiating adult axons depends not on preexisting nerve pathways but on unknown cues that appear to be associated with the disc epithelium itself (Blair et al. 1987).

Nonetheless, the nerves of the larva probably do perform an essential guidance function in providing a bridge between the developing discs and the differentiating adult central nervous system at metamorphosis. This guidance function has been amply documented for the larval visual nerve (Bolwig’s nerve) and the axons of adult retinula cells that grow along it, and other discs are closely associated with larval nerves that link them to the central nervous system (Bate, 1978). It is only in the case of the leg that the axons of the larval sensory cells happen to be included within the disc, rather than running close to it and thus provide a preexisting pathway for adult axons to grow along inside the disc as well as onwards from the base of the disc to the central nervous system.

As demonstrated in the case of two essentially different structures, the optic lobes of the brain (Tix et al. 1989) and the leg imaginal discs, selective staining of cells with a low number of cell divisions during postembryonic development uncovers preexisting larval neurons in developing imaginal structures of Drosophila and possibly other holometabolous insects. Since a lineage tracer injected into the embryo before cellularization labels the entire larval cell lineage, and given that the tracer becomes significantly diluted in all cells of the imaginal lineage, the method potentially provides a way of mapping completely the persisting larval neuronal scaffold in the fully developed adult central and peripheral nervous system.

We thank Jose Campos-Ortega for working facilities and comments on the manuscript, Eva Varus for expect technical assistance and Alberto Ferrus for providing MAb 22C10. This work was supported by grants from the Deutsche Forschungs-gemeinschaft to G.M.T. and from the SERC to M.B.

Bastiani
,
M. J.
,
Doe
,
C. Q.
,
Helfand
,
S. L.
and
Goodman
,
C. S.
(
1985
).
Neuronal specificity and growth cone guidance in grasshopper and Drosophila embryos
.
Trends Neurosci
.
8
,
257
266
.
Bate
,
C. M.
(
1976
).
Pioneer neurones in an insect embryo
.
Nature
260
,
54
56
.
Bate
,
C. M.
(
1978
).
Development of sensory systems
.
In Handbook of Sensory Physiology
(ed.
M.
Jacobson
), Vol.
IX
, pp.
1
53
.
Springer-Verlag
,
New York
.
Bentley
,
D.
and
Caddy
,
M.
(
1983
).
Pioneer axons lose directed growth after selective killing of guidepost cells
.
Nature
304
,
62
65
.
Berlot
,
J.
and
Goodman
,
C. S.
(
1984
).
Guidance of peripheral pioneer neurones in the grasshopper: Adhesive hierarchy of epithelial and neuronal surfaces
.
Science
223
,
493
496
.
Blair
,
S. S.
,
Murray
,
M. A.
and
Palka
,
J.
(
1987
).
The guidance of axons from transplanted neurons through aneural Drosophila wings
.
J. Neurosci
.
7
,
4165
4175
.
Canal
,
I.
and
Ferrus
,
A.
(
1986
).
The pattern of early neuronal differentiation in Drosophila melanogaster
.
J. Neurogenet
.
3
,
293
319
.
Fujita
,
S. C.
,
Zipursky
,
S. L.
,
Benzer
,
S.
,
Ferrus
,
A.
and
Shotwell
,
S. L.
(
1982
).
Monoclonal antibodies against Drosophila nervous system
.
Proc. natn. Acad. Sci. U.S.A
.
79
,
7929
7933
.
Jan
,
J. N.
,
Ghysen
,
A.
,
Christoph
,
I.
,
Barbel
,
S.
and
Jan
,
L. Y.
(
1985
).
Formation of neuronal pathways in the imaginai discs of Drosophila melanogaster
.
J. Neurosci.
5
,
2453
2464
.
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1982
).
Antibodies to horseradish peroxidase as specific neuronal markers in Drosophila and grasshopper embryos
.
Proc. natn. Acad. Sci. U.S.A
.
79
,
2700
2704
.
Keilin
,
D.
(
1915
).
Recherches sur les larves Dipteres cyclorrhaphes
.
Bull. Sci. Fr. Belg
.
49
,
15
198
.
Kuwada
,
J. Y.
(
1986
).
Cell recognition by neuronal growth cones in a simple vertebrate embryo
.
Science
233
,
740
746
.
Murray
,
M.
,
Schubiger
,
M.
and
Palka
,
J.
(
1984
).
Neuron differentiation and axon growth in the developing wing of Drosophila melanogaster
.
Devi Biol
.
104
,
259
273
.
Palka
,
J.
(
1986
).
Neurogenesis and axonal pathfinding in invertebrates
.
Trends Neurosci
.
9
,
482
485
.
Poodry
,
C. A.
and
Schneiderman
,
H. A.
(
1970
).
The ultrastructure of the developing leg of Drosophila melanogaster
.
Wilhelm Roux’s Arch. Entw Mech. Org
.
166
,
1
44
.
Schubiger
,
G.
(
1968
).
Anlageplan, Determinationszustand und Transdeterminationsleistungen der männlichen Vorderbeinscheibe von Drosophila melanogaster
.
Wilhelm. Roux’s Arch. EntwMech. Org
.
160
,
9
40
.
Technau
,
G. M.
(
1986
).
Lineage analysis of transplanted individual cells in embryos of Drosophila melanogaster. I. The method
.
Roux’s Arch, devl Biol
.
195
,
389
398
.
Tix
,
S.
,
Minden
,
J. S.
and
Technau
,
G. M.
(
1989
).
Pre-existing neuronal pathways in the developing optic lobes of Drosophila
.
Development
105
,
739
746
.